chore: regenerate crapload baseline for latest gaze version#471
Conversation
The baseline had stale GazeCRAP scores (None/0) from an older gaze version. Regenerating updates all scores so CI no longer flags false regressions on functions like FormatScanSummary. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> Signed-off-by: Hannah Braswell <hbraswel@redhat.com>
hbraswelrh
force-pushed
the
chore/update-crapload-baseline
branch
from
8d19a16 to
d02f591
Compare
marcusburghardt
left a comment
marcusburghardt
left a comment
There was a problem hiding this comment.
LGTM — CI is green and the baseline regeneration correctly addresses the stale GazeCRAP scores.
One note for awareness: The new recommended_actions section contains 20 entries with absolute file paths from the generating machine (e.g., /Users/hbraswel/github/complytime/complyctl/cmd/...). The scores, worst_crap, and worst_gaze_crap sections correctly use relative paths.
This is a gaze upstream issue — the newer version emits absolute paths in recommended_actions because Score.File is populated from gocyclo.Stat.Pos.Filename (which uses the absolute walk root) and no relativization was applied before building the summary.
A fix has been proposed upstream in gaze to relativize all Score.File paths against the module root after computeScores() returns. Once that fix lands, regenerating the baseline will produce relative paths everywhere.
Not a blocker for this PR.
Line numbers shifted in cmd/behavioral-report/main.go after upstream main updates. Regenerated baseline to reflect current state. Signed-off-by: Hannah Braswell <hbraswelrh@gmail.com> Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
3 participants
Summary
.gaze/baseline.jsonto include up-to-date GazeCRAP scoresNone/0 for all functionsFormatScanSummaryGazeCRAP 0 → 9) on any PR, regardless of code changesTest plan
make crapload-checklocally and verify no regressions are reportedmake craploadlocally and verify the text output matches the updated baseline scores🤖 Generated with Claude Code